Gaurav Shandilya

SUMMARY

• Seasoned Machine Learning Expert with vast experience working on Scikit-learn, Natural Language Processing, Deep Learning (Re-enforcement learning, Principal Component Analysis, Computer Vision).

• Experience Professional working on Deep Learning frameworks – Tensorflow, Theano, Keras and PyTorch.

• 5 years of experience using Python for Data Analysis, Data scrubbing, Data aggregation, Data Management, Data normalization, Data standardization, Data Enrichment, Data manipulation, Data staging, Data Classification, Data Reporting at Mizuho, JP Morgan, Morgan Stanley.

• Domain Experience in Data Analysis, Data Analytics, Data Visualization using Tableau tool.

• Developed machine learning models relying on decision tress, random forest, logistic regression, Linear Regression, Supervised Vector Models (SVM), Decision Forest for Asset Pricing & Loans Recoveries Forecasting (CCRM).

• Good experience in developing Machine Learning algorithms - K-NN, Randome Forest, XGBoost, Ada Boost, K-NN, Naive Bayes, Clustering Classification and Time Series methods.

• Experience on Deep Learning Models - Recurring Neural Networks (RNN), Deep Neural Networks

(DNN), Auto Encoders, RBM- Boltzman Machines, AutoEncoders models for building Recommender Systems.

• Good experience in developing web applications implementing Model View Control architecture using Django, Flask and web applications framework.

• Knowledge about setting up Python REST API Framework using Django.

• Developed intricate algorithms based on deep dive statistical analysis and Predictive Data modeling that were used to deepen relationships, strengthen longevity and personalize interactions with Enterprise Data Warehouse (EDH).

• Experience working on OLTP and OLAP Data Modeling. Facts and Dimensions tables and Dimension Modeling (Star schema and Snow Flake schema)

• An advanced ability to interact on various database and file storage systems (Oracle, No SQL).

• Vast experience working with large dataset analysis using SAS, Python, R, PROC SQL.

• Experience in building, maintaining, multiple AWS EMR clusters of different sizes and configurations and setting up rack topology for large cluster also in Spark.

• Experience in deployment of Spark/Scala and Streaming integration with Cassandra.

• Used Amazon’s Simple Storage Service (S3), Amazon Elastic MapReduce (EMR).

• Experience with SPARK improving the performance and optimization of existing algorithms using SPARK Context, SPARK-SQL, Hive, DATA FRAME, PAIR RDD’S.

• Developed SPARK Code using SCALA and SPARK-SQL, Streaming using Kafka for faster processing of data.

• Experience with Python 3.6(Numpy, Pandas, Matplotlib, NLTK, spaCy and Sci-kit learn).

• Highly effective in building/communicating project requirements with stakeholders and executives and managing technical development teams.

• Self-starter with excellent presentation, communication, coordination, documentation, project planning and interpersonal skills.

• Ability to meet deadlines and handle pressure coordinating multiple tasks in project environment. TECHNICAL SKILLS

• Programming Languages: Python, Python Django, Python Flask, Java, Apache Spark, Scala, R, Node.js, C#, PhP, Golang. Familiarity with Tensorflow, Keras, PyTorch, Hadoop HDFS, AWS S3, GoGRPC.

• Experience with mathematical, statistical Python libraries such as pandas, scikit-learn, Numpy and Scipy, Tensorflow, Keras, MATLAB and Tableau Desktop & Server.

• Advanced mathematics and physics toolset knowledge of software best practices and applied machine learning ideally suited to tackle bleeding-edge challenges in AI and Deep Learning. PROFESSIONAL EXPERIENCE

Mizuho, NYC, NY

Data Scientist Oct 2016- Till Date

Project: This program involved building Trading BOT application supporting Traders to make Strategy decisions on Buy, Sell, Hold Strategies. Trader BOT allowed Traders to optimize the Portfolios with NLP Automation services leveraging Google Dialog Flow framework.

Responsibilities:

• Architect of new machine learning CCAR Loss Forecasting scoring model with ability to serve models developed in several different languages and frameworks.

• Actively involved in all phases of the data science project life cycle including data extraction, data cleansing, statistical modeling, and data visualization with large datasets of structured and unstructured data for Trader BOT application.

• Performed data analysis on datasets of Mortgage loans, Equities, Debt securities using Python 3 panda series numpy, matplotlib, scikit-learn. Converted Sas programs for risk scoring in to Python for AWS S3 migration.

• Developed Customer Churn Out Prediction model using Sklearn Logistic Regression Model. Leveraged dataset for Canceled Loans, Received Loans, Rejected Loans to compute Churn Value counts and performed Model accuracy, precision score, F1 Score, Recall Score.

• Used the sklearn.model_selection for Cross Validation Score to compute Churn Prediction accuracies based on the Classifier (Logistic Regression).

• Minimized the Churn of the Loans sanctioning of Customers by performing Data Cleansing, Correlation Plot, Correlation Matrix, One Hot Encoding, Feature Scaling & Balancing, Model Building and perform K-Fold Cross Validation.

• Developed Deep Learning Model with Tensorflow, performed data wrangling and data pipes to consume the data in the scoring model and contributed to Python robotics software stack.

• Implemented Python variants of various learning algorithms, such as Generalized Additive Models and Constrained Linear Models.

• Developed features and applications using Python and Django in test driven Development and pair based programming.

• Implemented Machine Learning models for Loans expected losses predictions using Deep Forest Search algorithm, K Near algorithm and Supervised Vector Models.

• Developed entire backend modules using Python on Django Web Framework.

• Responsible for using Machine Learning Models such as Linear Regression and Logistic Regression for predicting the Recoveries percentage of the Commercial Loans as part of Capital Loss Forecasting Models.

Bank of America, Jersey City, NJ

Data Scientist Oct 2013- Oct 2016

Project: Project: Bank of America is one of the world's largest financial institutions, serving individual consumers, small and middle market businesses and large corporations with a full range of banking, investing, asset management and other financial and risk-management products and services Responsibilities:

• Responsible for creating Data Frames on the Securities using Python panda series, used numpy operations and scipy to evaluate expressions used in Black-Scholes model for Option Pricing.

• Used Python Matplotlib for graphical representation of the histograms/pi-charts/bar charts

• Solidified and scaled end-to-end PySpark ETL machine learning pipeline, resulting in a ~5X increase to handle data scale and ~5X decrease of training time.

• Exclusively used AWS Lambda microservices to schedule API’s/Methods in a desired interval periods. Performed AWS S3 bucket connection with Snowflake for data validations for Cloud Migration of CCAR ALLM models supporting FRY-14M, FRY-14Q reporting requirements.

• Reduced feature engineering development times by 3X using univariate, multi-variate, Lasso and Regressor methods to detect independent features and automated the data pipeline

• Developed the Campaigns optimization using Thomson Sampling, Sci-kit learn Lasso Regression model and improved the Campaign returns ROI by 240%.

• Created Customer Life Time Value metric by assimilating marketing expense, mail campaigns, customer subscription. Performed A/B testing to understand impact of promotions.

• Deployed, debugged and maintained complex, distributed software stacks, containing Apache Spark, Hadoop HDFS, IPython Notebook servers, on cloud based AWS system.

• Developed production algorithms for feature analysis and selection. Provided Dashboard and automated reports for business stakeholders.

• Created Business Intelligence Reports us and Dashboard metrics using Tableau, BO Reports

• Was responsible for writing shell scripts automation of scripts to validate SQL code in Snowflake Vs Teradata for supporting Risk Capital Portfolio models to support AWS migration. JP Morgan Chase, NYC, NY

Business Intelligence Analyst October 2011- October 2013 Project: Liquidity Risk Infrastructure program is mission critical effort to enhance J P Morgan liquidity risk management monitoring and reporting capabilities. This program provided treasury details in to liquidity products, currencies, legal entities, regulatory jurisdiction. In addition platform provided stress testing, contingency funding plans to make appropriate funding decisions.

Responsibilities:

• Built the IBM Business Intelligence Cognos Report by writing mapping specifications to reconcile Transaction Position level balances to PeopleSoft Corporate General Ledger account(s).

• Performed Data analysis using complex sql joins for Fixed Income, OTC Derivatives, Collateral, Securities Finance & Commodities for processing platforms- IMPACT, GENESIS, SCPP, NAPOLI.

• Performed data analysis using Python 2 on the data frames derived on large datasets of Bond Valuations and plotted results of regression analysis.

• Responsible for migration of Ledger for SCPP system to new SAP account. Reconciled all Balance Sheet balances in spread sheet using Macros such as Notional, Current Market Value, Present Value, Face Value, Un-amortize Amount, Profit & Loss to Oracle General Ledger.

• Daily basis ran the reconciliation report for migration of accounts and reported known differences between original and new Ledger balances.

• Built a complex Enterprise Java ecosystem in collaboration with development team.

• Performed Python Snowflake connection to test the data validations between various source/target systems supporting Liquidity Risk measures used for Stress Testing.

• Responsible for reporting Performance KPIs via Data Visualization of the score card using Tableau analytics.

Bank of Tokyo Jersey City, NJ

Business Intelligence Analyst October 2010-October 2011 Project: Bank of Tokyo is one of the world's largest financial institutions, serving individual consumers, small and middle market businesses and large corporations with a full range of banking, investing, risk-management products and services. Project was to replace existing Fixed Income processing system-IMPACT with new system-Genesis for Trade Capture, Settlement, Downstream Reporting-Regulatory Reporting and other Risk Management Reporting services.

Responsibilities:

• Analyzed Reference Data for Fixed Income products (CMBS, MBS, ABS, Bonds, Tbills, Repos, Collateral) from various feeder systems like BLOOMBERG.

• Performed data profiling and data mapping of Reference data (Client Master Reference and Security Master Reference data) for source to target by loading data first in to STAGING then DWH Fixed Income Data Mart.

• As a Lead Data Analyst worked on SMF (Security Master File) that comprising analyzing feeds for ADS (Asset Data Services) like Bloomberg Indicative, Bloomberg Per Security Price/Yield/ Duration, Bloomberg Bulk-Rates. Performed File Loading/Monitoring, Data Scrubbing/Monitoring/Exception.

• Was responsible for designing the Customer Relationship Management System and integrating of Customer data from various platforms in to the (Goldtier).

• Analyzed Fixed Income reference data, Indicative data and designed data model consisting of Security Fact, Dimensions such as GL dim, Cost center dim, Customer dim.

• Using ETL tool performed aggregations/lookups/update strategy transformations on Reference data such as Ratings, Market Price Information, Counterparty/Custodian/Broker/Dealer Information.

• Was responsible for writing SQL code for QA Reports supporting reconciliation of Security Master

& Client Reference data for Exception Management and matching sub-ledger to Oracle Ledger

(Chart of Accounts) -O300, T300.

Morgan Stanley, NY

Data Analyst July 2006 – March 2010

Project: Morgan Stanley provides investment management, retail and commercial banking, consumer finance, and investment banking products and services to individuals and companies throughout the United States and, for certain businesses, internationally. Morgan Stanley Prime Brokerage provides hedge fund clients with stable and prudent financing, allowing for the achievement of investment objectives in accordance with regulatory and risk-based margin requirements.

Responsibilities:

• Ensured continual improvement of analytical approaches for pricing, trading, risk management through quant models within financial engineering framework for Morgan Stanley’s Global Asset Servicing Platform.

• Analyzed Reference data and Indicative data/Static data from various feeds to create system workflow for Fixed Income (CMO, MBS, ABS).

• Worked as a Lead Analyst to design the data acquisition strategy for Reference Data to support Asset Servicing by housing Reference data in data warehouse.

• Responsible to build the data movement for downstream consumers from SMF(Security Master File) to EDB(Enterprise Database), EDB to OMS(Long View Trading System), EDB to TAX.

• Was responsible for writing requirements for complex Data Stage of Holding &Transactions reference data in Teradata region of Enterprise Data Warehouse.

• Developed User and Support documentation, Test plans and use cases as a basis for performing Integration and System testing for User Acceptance Testing (UAT) Brown Brothers Harriman, Boston, MA

Role: Data Analyst DEC 2004 – July2006

Project: Action World2.0 Project was to automate and re-engineer processing of Re-org actions which were processed manually in Action World1.0. The focus of Action World 2.0 was to build upon existing capabilities in Action World1.0 and significantly improve BBH’s ability to further differentiate its services offering to its existing and new clients in the future as a Global Custodian

Responsibilities:

• Interacted with clients to asses needs, identify key challenges, define project scope and deliverables for BBH Global Custody Services.

• Assess users' needs to determine functional and business requirements for Options Trading application.

• Wrote the Functional specifications, technical specifications and Business Rules for the, Stock Loan/Stock Borrower system reduced manual effort required to process Stock Loan payments.

• Prioritized project efforts, developed project plans, process flows, Use Cases and SRS for various assembly line of Corporate Actions such as Notification, Entitlement Processing and Settlement of Payments.

• Worked with Business Requirement document and User Centered Design (UCD) to create UML use case, class and sequence diagrams.

• Lead data analysis review sessions of Sourcing Stock Loan feed, created process flowcharts for reference data, financial instruments data attributes and security master to include CUSIP; diagrams and static data mapping with MS Visio.

• Developed Test plans and use cases as a basis for performing Integration and System Testing for User acceptance Testing.

EDUCATION

MS in Computer Science

NJIT (New Jersey Institute of Technology), NJ

Contact this candidate